Crack propagation in thermal induced stress field

[Lodham]

Hi there,

I have performed a coupled temp-displacement analysis to simulate the cooling stage of two bodies in contact.
Due to the thermal expansion mismatch, a stress tensile stress field appears in one of the two bodies.

Now I would like to study how a crack would initiate and propagate in this body with tensile stresses.
For this, I was planning to use the XFEM capabilities of Abaqus V6.11. However, it seems that crack propagation cannot be done in a coupled Temp-displacement analysis.

The solution would be to perform the Coupled T°/displ analysis and then to perform a crack propagation modelin in a standard analysis. However I cannot do this in two steps, since I would like to add the crack only in the second step.
Can I do it by running two different simulations?
Can I export the geometry and stress field of my first simulation and use it as input data for a second simulation in which I would add a Maxps damage criterion to propagate an XFEM crack?

Thank you for your help
Pierre

 

[corus]

Importing the stress field would serve no purpose as it would be the stress field without a crack, presumably.

If you have the temperatures in the odb or fil file then import them into another model with the crack. There is an option that would interpolate the temperatures based upon the co-ordinates rather than the node numbering, whcih would allow this.

 

[Lodham]

The thing is the parts are loaded without any crack at the beginning.
Once the stress induced by the thermal cooling has reached a certain critical value, a crack initiates and propagates dynamically.

Now, I need to import a stress field (or temperature field), but also the deformed geometry coming from the previous coupled Temp-Displ analysis.
I have never done this before.
Do I need to use a restart analysis?
How can I export the final deformed geometry in order to import it with the associated stress field in a new analysis?

Thanks again
Pierre

 

[Lodham]

OK I found how to import the deformed mesh as well as the stress field.
You have to be careful because Abaqus add a "-1" extension to the part you import, and the name of the part does not correspond anymore to the one of the .odb file that you have imported.
You thus have to delete this "-1" extension in the "Assembly" Module.

Now the crack initiation is correct but I would definitely need a coupled Temp-Displacement analysis with my xfem crack propagation to get good results.
This does not seem to be possible yet :'(

Pierre

 

[mikechy]

Hi Pierre,
I have a similar problem, but in the previous step my model was plastically deformed. Thus I also cannot run the next step directly, but must run restart instead of it. For some reasons, XFEM does not work with restart (there is no XFEM output in odb result file). Can you send me your input decks as templates to find, where I made a mistake? Thanks in advance.
Regards,
Mike